platform/kernel/linux-starfive.git
14 years agojump label: Make text_poke_early() globally visible
Jason Baron [Fri, 17 Sep 2010 15:08:56 +0000 (11:08 -0400)]
jump label: Make text_poke_early() globally visible

Make text_poke_early available outside of alternative.c. The jump label
patchset wants to make use of it in order to set up the optimal no-op
sequences at run-time.

Signed-off-by: Jason Baron <jbaron@redhat.com>
LKML-Reference: <04cfddf2ba77bcabfc3e524f1849d871d6a1cf9d.1284733808.git.jbaron@redhat.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
14 years agojump label: Make dynamic no-op selection available outside of ftrace
Jason Baron [Fri, 17 Sep 2010 15:08:51 +0000 (11:08 -0400)]
jump label: Make dynamic no-op selection available outside of ftrace

Move Steve's code for finding the best 5-byte no-op from ftrace.c to
alternative.c. The idea is that other consumers (in this case jump label)
want to make use of that code.

Signed-off-by: Jason Baron <jbaron@redhat.com>
LKML-Reference: <96259ae74172dcac99c0020c249743c523a92e18.1284733808.git.jbaron@redhat.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
14 years agoperf: Undo the per cpu-context timer stuff
Peter Zijlstra [Fri, 17 Sep 2010 09:28:50 +0000 (11:28 +0200)]
perf: Undo the per cpu-context timer stuff

Revert the timer per cpu-context timers because of unfortunate
nohz interaction. Fixing that would have been somewhat ugly, so
go back to driving things from the regular tick. Provide a
jiffies interval feature for people who want slower rotations.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Yinghai Lu <yinghai@kernel.org>
LKML-Reference: <20100917093009.519845633@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf: Fix perf_event_exit_cpu_context()
Peter Zijlstra [Fri, 17 Sep 2010 09:28:49 +0000 (11:28 +0200)]
perf: Fix perf_event_exit_cpu_context()

Use the right cpu-context.. spotted by preempt warning on
hot-unplug

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
Cc: Robert Richter <robert.richter@amd.com>
LKML-Reference: <20100917093009.461794357@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf: Complete software pmu grouping
Peter Zijlstra [Fri, 17 Sep 2010 09:28:48 +0000 (11:28 +0200)]
perf: Complete software pmu grouping

Aside from allowing software events into a !software group,
allow adding !software events to pure software groups.

Once we've moved the software group and attached the first
!software event, the group will no longer be a pure software
group and hence no longer be eligible for movement, at which
point the straight ctx comparison is correct again.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <20100917093009.410784731@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf_events: Fix broken event grouping
Stephane Eranian [Fri, 17 Sep 2010 09:28:47 +0000 (11:28 +0200)]
perf_events: Fix broken event grouping

Events were not grouped anymore. The reason was that in
perf_event_open(), the field event->group_leader was
initialized before the function looked up the group_fd
to find the event leader. This patch fixes this by
reordering the code correctly.

Signed-off-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Robert Richter <robert.richter@amd.com>
LKML-Reference: <20100917093009.360420946@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agotracing, perf: Add more power related events
Jean Pihet [Tue, 7 Sep 2010 07:21:32 +0000 (09:21 +0200)]
tracing, perf: Add more power related events

This patch adds new generic events for dynamic power management
tracing:

 - clock events class: used for clock enable/disable and for
   clock rate change,
 - power_domain events class: used for power domains transitions.

The OMAP architecture will be using the new events for PM debugging,
however the new events are made generic enough to be used by all
platforms.

Signed-off-by: Jean Pihet <j-pihet@ti.com>
Acked-by: Thomas Renninger <trenn@suse.de>
Cc: discuss@lesswatts.org
Cc: linux-pm@lists.linux-foundation.org
Cc: Thomas Renninger <trenn@suse.de>
Cc: Arjan van de Ven <arjan@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Len Brown <len.brown@intel.com>
Cc: Kevin Hilman <khilman@deeprootsystems.com>
LKML-Reference: <AANLkTinUmbSUUuxUzc8++pcb9gd1CZFdyTQFrveTBXyV@mail.gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agokprobes: Add sparse context annotations
Namhyung Kim [Wed, 15 Sep 2010 01:04:30 +0000 (10:04 +0900)]
kprobes: Add sparse context annotations

This removes following warnings when build with C=1

 warning: context imbalance in 'kretprobe_hash_lock' - wrong count at exit
 warning: context imbalance in 'kretprobe_table_lock' - wrong count at exit
 warning: context imbalance in 'kretprobe_hash_unlock' - unexpected unlock
 warning: context imbalance in 'kretprobe_table_unlock' - unexpected unlock

Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Acked-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
LKML-Reference: <1284512670-2369-6-git-send-email-namhyung@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agokprobes: Remove __dummy_buf
Namhyung Kim [Wed, 15 Sep 2010 01:04:29 +0000 (10:04 +0900)]
kprobes: Remove __dummy_buf

Remove __dummy_buf which is needed for kallsyms_lookup only.
use kallsysm_lookup_size_offset instead.

Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Acked-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
LKML-Reference: <1284512670-2369-5-git-send-email-namhyung@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agokprobes: Make functions static
Namhyung Kim [Wed, 15 Sep 2010 01:04:28 +0000 (10:04 +0900)]
kprobes: Make functions static

Make following (internal) functions static to make sparse
happier :-)

 * get_optimized_kprobe: only called from static functions
 * kretprobe_table_unlock: _lock function is static
 * kprobes_optinsn_template_holder: never called but holding asm code

Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Acked-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
LKML-Reference: <1284512670-2369-4-git-send-email-namhyung@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agokprobes: Verify jprobe entry point
Namhyung Kim [Wed, 15 Sep 2010 01:04:27 +0000 (10:04 +0900)]
kprobes: Verify jprobe entry point

Verify jprobe's entry point is a function entry point
using kallsyms' offset value.

Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Acked-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
LKML-Reference: <1284512670-2369-3-git-send-email-namhyung@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agokprobes: Remove redundant address check
Namhyung Kim [Wed, 15 Sep 2010 01:04:26 +0000 (10:04 +0900)]
kprobes: Remove redundant address check

Remove call to kernel_text_address() in register_jprobes()
because it is called right after in register_kprobe().

Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Acked-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
LKML-Reference: <1284512670-2369-2-git-send-email-namhyung@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf events: Clean up pid passing
Matt Helsley [Mon, 13 Sep 2010 20:01:20 +0000 (13:01 -0700)]
perf events: Clean up pid passing

The kernel perf event creation path shouldn't use find_task_by_vpid()
because a vpid exists in a specific namespace. find_task_by_vpid() uses
current's pid namespace which isn't always the correct namespace to use
for the vpid in all the places perf_event_create_kernel_counter() (and
thus find_get_context()) is called.

The goal is to clean up pid namespace handling and prevent bugs like:

https://bugzilla.kernel.org/show_bug.cgi?id=17281

Instead of using pids switch find_get_context() to use task struct
pointers directly. The syscall is responsible for resolving the pid to
a task struct. This moves the pid namespace resolution into the syscall
much like every other syscall that takes pid parameters.

Signed-off-by: Matt Helsley <matthltc@us.ibm.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Robin Green <greenrd@greenrd.org>
Cc: Prasad <prasad@linux.vnet.ibm.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Mahesh Salgaonkar <mahesh@linux.vnet.ibm.com>
LKML-Reference: <a134e5e392ab0204961fd1a62c84a222bf5874a9.1284407763.git.matthltc@us.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf events: Split out task search into helper
Matt Helsley [Mon, 13 Sep 2010 20:01:19 +0000 (13:01 -0700)]
perf events: Split out task search into helper

Split out the code which searches for non-exiting tasks into its own
helper. Creating this helper not only makes the code slightly more
readable it prepares to move the search out of find_get_context() in
a subsequent commit.

Signed-off-by: Matt Helsley <matthltc@us.ibm.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Robin Green <greenrd@greenrd.org>
Cc: Prasad <prasad@linux.vnet.ibm.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Mahesh Salgaonkar <mahesh@linux.vnet.ibm.com>
LKML-Reference: <561205417b450b8a4bf7488374541d64b4690431.1284407762.git.matthltc@us.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agohw breakpoints: Fix pid namespace bug
Matt Helsley [Mon, 13 Sep 2010 20:01:18 +0000 (13:01 -0700)]
hw breakpoints: Fix pid namespace bug

Hardware breakpoints can't be registered within pid namespaces
because tsk->pid is passed rather than the pid in the current
namespace.

(See https://bugzilla.kernel.org/show_bug.cgi?id=17281 )

This is a quick fix demonstrating the problem but is not the
best method of solving the problem since passing pids internally
is not the best way to avoid pid namespace bugs. Subsequent patches
will show a better solution.

Much thanks to Frederic Weisbecker <fweisbec@gmail.com> for doing the
bulk of the work finding this bug.

Signed-off-by: Matt Helsley <matthltc@us.ibm.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Robin Green <greenrd@greenrd.org>
Cc: Prasad <prasad@linux.vnet.ibm.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Mahesh Salgaonkar <mahesh@linux.vnet.ibm.com>
LKML-Reference: <f63454af09fb1915717251570423eb9ddd338340.1284407762.git.matthltc@us.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf, sh: Fix patch merge foul-up
Matt Fleming [Mon, 13 Sep 2010 19:17:29 +0000 (20:17 +0100)]
perf, sh: Fix patch merge foul-up

3f6da390 ("perf: Rework and fix the arch CPU-hotplug hooks") introduced
this breakage. sh_pmu_setup() is missing an opening curly brace.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <20100913191729.GA6440@console-pimps.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoalpha: Fix HW performance counters to be stopped properly
Michael Cree [Sun, 12 Sep 2010 05:37:24 +0000 (17:37 +1200)]
alpha: Fix HW performance counters to be stopped properly

Also fix a few compile errors due to undefined and duplicated
variables.

Signed-off-by: Michael Cree <mcree@orcon.net.nz>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1284269844-23251-1-git-send-email-mcree@orcon.net.nz>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agowatchdog: Avoid kernel crash when disabling watchdog
Stephane Eranian [Tue, 14 Sep 2010 13:34:01 +0000 (15:34 +0200)]
watchdog: Avoid kernel crash when disabling watchdog

In case you boot with the watchdog disabled, i.e., nowatchdog, then,
if you try to disable it via /proc/sys/kernel/watchdog, you get
a kernel crash. The reason is that you are trying to cancel a hrtimer
which has never been initialized.

This patch fixes this by skipping execution of
watchdog_disable_all_cpus() when the watchdog is marked
disabled from boot.

Signed-off-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <4c8f7a23.cae9d80a.2c11.0bb4@mx.google.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoMerge branch 'tip/perf/core' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt...
Ingo Molnar [Wed, 15 Sep 2010 08:27:31 +0000 (10:27 +0200)]
Merge branch 'tip/perf/core' of git://git./linux/kernel/git/rostedt/linux-2.6-trace into perf/core

14 years agotracing: Remove leftover FTRACE_ENABLE/DISABLE_MCOUNT enums
Steven Rostedt [Wed, 15 Sep 2010 02:19:46 +0000 (22:19 -0400)]
tracing: Remove leftover FTRACE_ENABLE/DISABLE_MCOUNT enums

The enums for FTRACE_ENABLE_MCOUNT and FTRACE_DISABLE_MCOUNT were
used as commands to ftrace_run_update_code(). But these commands
were used by the old nasty ftrace daemon that has long been slain.

This is a clean up patch to remove the references to these enums
and simplify the code a little.

Reported-by: Wu Zhangjin <wuzhangjin@gmail.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
14 years agotracing: Do not trace in irq when funcgraph-irq option is zero
Steven Rostedt [Tue, 14 Sep 2010 22:58:33 +0000 (18:58 -0400)]
tracing: Do not trace in irq when funcgraph-irq option is zero

When the function graph tracer funcgraph-irq option is zero, disable
tracing in IRQs. This makes the option have two effects.

1) When reading the trace file, do not display the functions that
   happen in interrupt context (when detected)

2) [*new*] When recording a trace, skip those that are detected
   to be in interrupt by the 'in_irq()' function

Note, in_irq() is updated at irq_enter() and irq_exit(). There are
still functions that are recorded by the function graph tracer that
is in interrupt context but outside the irq_enter/exit() routines.

Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
14 years agotracing: Add funcgraph-irq option for function graph tracer.
Jiri Olsa [Tue, 7 Sep 2010 14:53:44 +0000 (16:53 +0200)]
tracing: Add funcgraph-irq option for function graph tracer.

It's handy to be able to disable the irq related output
and not to have to jump over each irq related code, when
you have no interrest in it.

The option is by default enabled, so there's no change to
current behaviour. It affects only the final output, so all
the irq related data stay in the ring buffer.

Signed-off-by: Jiri Olsa <jolsa@redhat.com>
LKML-Reference: <20100907145344.GC1912@jolsa.brq.redhat.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
14 years agotracing: Fix reading of set_ftrace_filter across lists
Steven Rostedt [Tue, 14 Sep 2010 15:21:11 +0000 (11:21 -0400)]
tracing: Fix reading of set_ftrace_filter across lists

If we do:

 # cd /sys/kernel/debug
 # echo 'do_IRQ:traceon schedule:traceon sys_write:traceon' > \
    set_ftrace_filter
 # cat set_ftrace_filter

We get the following output:

 #### all functions enabled ####
 sys_write:traceon:unlimited
 schedule:traceon:unlimited
 do_IRQ:traceon:unlimited

This outputs two lists. One is the fact that all functions are
currently enabled for function tracing, the other has three probed
functions, which happen to have 'traceon' as their commands.

Currently, when reading the first list (functions enabled) the
seq_file code will receive a "NULL" from the t_next() function
causing it to exit early. This makes "read()" from userspace stop
reading the code at this boarder. Although read is allowed to do this,
some (broken) applications might consider this an end of file and
stop early.

This patch adds the start of the second list to t_next() when it
finishes the first list. It is a simple change and gives the
set_ftrace_filter file nicer reading ability.

Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
14 years agotracing: Keep track of set_ftrace_filter position and allow lseek again
Steven Rostedt [Fri, 10 Sep 2010 15:47:43 +0000 (11:47 -0400)]
tracing: Keep track of set_ftrace_filter position and allow lseek again

This patch keeps track of the index within the elements of
set_ftrace_filter and if the position goes backwards, it nicely
resets and starts from the beginning again.

This allows for lseek and pread to work properly now.

Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
14 years agotracing: Replace typecasted void pointer in set_ftrace_filter code
Steven Rostedt [Thu, 9 Sep 2010 14:00:28 +0000 (10:00 -0400)]
tracing: Replace typecasted void pointer in set_ftrace_filter code

The set_ftrace_filter uses seq_file and reads from two lists. The
pointer returned by t_next() can either be of type struct dyn_ftrace
or struct ftrace_func_probe. If there is a bug (there was one)
the wrong pointer may be used and the reference can cause an oops.

This patch makes t_next() and friends only return the iterator structure
which now has a pointer of type struct dyn_ftrace and struct
ftrace_func_probe. The t_show() can now test if the pointer is NULL or
not and if the pointer exists, it is guaranteed to be of the correct type.

Now if there's a bug, only wrong data will be shown but not an oops.

Cc: Chris Wright <chrisw@sous-sol.org>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
14 years agotracing: Do not reset *pos in set_ftrace_filter
Steven Rostedt [Thu, 9 Sep 2010 12:43:22 +0000 (08:43 -0400)]
tracing: Do not reset *pos in set_ftrace_filter

After the filtered functions are read, the probed functions are read
from the hash in set_ftrace_filter. When the hashed probed functions
are read, the *pos passed in is reset. Instead of modifying the pos
given to the read function, just record the pos where the filtered
functions ended and subtract from that.

Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
14 years agoMerge branch 'sched/urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/tip...
Linus Torvalds [Mon, 13 Sep 2010 21:34:27 +0000 (14:34 -0700)]
Merge branch 'sched/urgent' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'sched/urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  sched: Improve latencies under load by decreasing minimum scheduling granularity

14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux...
Linus Torvalds [Mon, 13 Sep 2010 19:51:22 +0000 (12:51 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/geert/linux-m68k

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k:
  m68k,m68knommu: Wire up fanotify_init, fanotify_mark, and prlimit64

14 years agoMerge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6
Linus Torvalds [Mon, 13 Sep 2010 19:49:55 +0000 (12:49 -0700)]
Merge branch 'release' of git://git./linux/kernel/git/aegl/linux-2.6

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6:
  [IA64] fix siglock

Quoth Tony:

 "I committed the fix for this last week prior to your -rc4 announcement
  reminding us to give proper "Reported-by:" credit.  This one should have
  had:

Reported-by: Tony Ernst <tee@sgi.com>
  and also

Much-useful-investigation-and-tracing-by: Hedi Berriche <hedi@sgi.com>
Much-useful-investigation-and-tracing-by: Petr Tesarik <ptesarik@novell.com>"
14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/sfrench/cifs-2.6
Linus Torvalds [Mon, 13 Sep 2010 19:47:08 +0000 (12:47 -0700)]
Merge git://git./linux/kernel/git/sfrench/cifs-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/sfrench/cifs-2.6:
  cifs: prevent possible memory corruption in cifs_demultiplex_thread
  cifs: eliminate some more premature cifsd exits
  cifs: prevent cifsd from exiting prematurely
  [CIFS] ntlmv2/ntlmssp remove-unused-function CalcNTLMv2_partial_mac_key
  cifs: eliminate redundant xdev check in cifs_rename
  Revert "[CIFS] Fix ntlmv2 auth with ntlmssp"
  Revert "missing changes during ntlmv2/ntlmssp auth and sign"
  Revert "Eliminate sparse warning - bad constant expression"
  Revert "[CIFS] Eliminate unused variable warning"

14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh...
Linus Torvalds [Mon, 13 Sep 2010 19:46:35 +0000 (12:46 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/ericvh/v9fs

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs:
  fs/9p: Don't use dotl version of mknod for dotu inode operations
  fs/9p: Use the correct dentry operations
  9p: Check for NULL fid in v9fs_dir_release()
  fs/9p: Fix error handling in v9fs_get_sb
  fs/9p, net/9p: memory leak fixes

14 years agoMerge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux...
Linus Torvalds [Mon, 13 Sep 2010 19:46:09 +0000 (12:46 -0700)]
Merge branch 'for_linus' of git://git./linux/kernel/git/jack/linux-fs-2.6

* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs-2.6:
  dquot: do full inode dirty in allocating space

14 years agoMerge branch 'next-spi' of git://git.secretlab.ca/git/linux-2.6
Linus Torvalds [Mon, 13 Sep 2010 19:45:50 +0000 (12:45 -0700)]
Merge branch 'next-spi' of git://git.secretlab.ca/git/linux-2.6

* 'next-spi' of git://git.secretlab.ca/git/linux-2.6:
  spi/pl022: move probe call to subsys_initcall()
  powerpc/5200: mpc52xx_uart.c: Add of_node_put to avoid memory leak
  spi/pl022: fix APB pclk power regression on U300
  spi/spi_s3c64xx: Warn if PIO transfers time out
  spi/s3c64xx: Fix incorrect reuse of 'val' local variable.
  spi/s3c64xx: Fix compilation warning
  spi/dw_spi: clean the cs_control code
  spi/dw_spi: Allow interrupt sharing
  spi/spi_s3c64xx: Increase dead reckoning time in wait_for_xfer()
  spi/spi_s3c64xx: Move to subsys_initcall()
  spi: free children in spi_unregister_master, not siblings
  gpiolib: Add 'struct gpio_chip' forward declaration for !GPIOLIB case
  of: Fix missing includes - ll_temac
  spi/spi_s3c64xx: Staticise non-exported functions
  spi/spi_s3c64xx: Make probe more robust against missing board config

14 years agom68k,m68knommu: Wire up fanotify_init, fanotify_mark, and prlimit64
Geert Uytterhoeven [Mon, 16 Aug 2010 12:05:20 +0000 (14:05 +0200)]
m68k,m68knommu: Wire up fanotify_init, fanotify_mark, and prlimit64

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Acked-by: Greg Ungerer <gerg@uclinux.org>
14 years agosched: Improve latencies under load by decreasing minimum scheduling granularity
Ingo Molnar [Sun, 12 Sep 2010 06:14:52 +0000 (08:14 +0200)]
sched: Improve latencies under load by decreasing minimum scheduling granularity

Mathieu reported bad latencies with make -j10 kind of kbuild
workloads - which is mostly caused by us scheduling with a
too coarse granularity.

Reduce the minimum granularity some more, to make sure we
can meet the latency target.

I got the following results (make -j10 kbuild load, average of 3
runs):

 vanilla:

  maximum latency: 38278.9 Âµs
  average latency:  7730.1 Âµs

 patched:

  maximum latency: 22702.1 Âµs
  average latency:  6684.8 Âµs

Mathieu also measured it:

|
| * wakeup-latency.c (SIGEV_THREAD) with make -j10
|
| - Mainline 2.6.35.2 kernel
|
| maximum latency: 45762.1 Âµs
| average latency: 7348.6 Âµs
|
| - With only Peter's smaller min_gran (shown below):
|
| maximum latency: 29100.6 Âµs
| average latency: 6684.1 Âµs
|

Reported-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Suggested-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <AANLkTi=8m4g01wZPacySoF7U0PevTNVgJoZZrHiUD-pN@mail.gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf ui browser: Don't use windows, slang is enough
Arnaldo Carvalho de Melo [Mon, 13 Sep 2010 13:25:04 +0000 (10:25 -0300)]
perf ui browser: Don't use windows, slang is enough

They are useless and take away precious columns and lines, so stop using
windows.

One more step in removing newt code, that after all is not being useful
at all for the coalescing TUI model in perf.

Suggested-by: Christoph Hellwig <hch@infradead.org>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Tom Zanussi <tzanussi@gmail.com>
LKML-Reference: <20100822082003.GB7365@infradead.org>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
14 years agoperf: Fix free_event()
Peter Zijlstra [Mon, 13 Sep 2010 09:15:58 +0000 (11:15 +0200)]
perf: Fix free_event()

With the context rework stuff we can actually end up freeing an event
before it gets attached to a context.

Reported-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf: Sanitize the RCU logic
Peter Zijlstra [Mon, 13 Sep 2010 09:06:55 +0000 (11:06 +0200)]
perf: Sanitize the RCU logic

Simplify things and simply synchronize against two RCU variants for
PMU unregister -- we don't care about performance, its module unload
if anything.

Reported-by: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agofs/9p: Don't use dotl version of mknod for dotu inode operations
Aneesh Kumar K.V [Mon, 30 Aug 2010 17:43:07 +0000 (17:43 +0000)]
fs/9p: Don't use dotl version of mknod for dotu inode operations

We should not use dotlversion for the dotu inode operations

Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
14 years agofs/9p: Use the correct dentry operations
Aneesh Kumar K.V [Mon, 30 Aug 2010 16:04:35 +0000 (16:04 +0000)]
fs/9p: Use the correct dentry operations

We should use the cached dentry operation only if caching mode is enabled

Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
14 years ago9p: Check for NULL fid in v9fs_dir_release()
jvrao [Wed, 25 Aug 2010 16:26:21 +0000 (16:26 +0000)]
9p: Check for NULL fid in v9fs_dir_release()

NULL fid should be handled in cases where we endup calling v9fs_dir_release()
before even we instantiate the fid in filp.

Signed-off-by: Venkateswararao Jujjuri <jvrao@linux.vnet.ibm.com>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
14 years agofs/9p: Fix error handling in v9fs_get_sb
Aneesh Kumar K.V [Tue, 24 Aug 2010 10:30:49 +0000 (10:30 +0000)]
fs/9p: Fix error handling in v9fs_get_sb

This was introduced by 7cadb63d58a932041afa3f957d5cbb6ce69dcee5

Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
14 years agofs/9p, net/9p: memory leak fixes
Latchesar Ionkov [Tue, 24 Aug 2010 18:13:59 +0000 (18:13 +0000)]
fs/9p, net/9p: memory leak fixes

Four memory leak fixes in the 9P code.

Signed-off-by: Latchesar Ionkov <lucho@ionkov.net>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
14 years agoperf_events: Fix BTS interrupt handling to avoid being dazed by NMI (v2)
Stephane Eranian [Fri, 10 Sep 2010 11:28:01 +0000 (13:28 +0200)]
perf_events: Fix BTS interrupt handling to avoid being dazed by NMI (v2)

Fix a bug introduced with commit de725de and the change in the
meaning of the return value of intel_pmu_handle_irq(). With the
current code, when you are using the BTS, you get 'dazed by NMI'
each time the BTS buffer fills up.

BTS does interrupt on the PMU vector, thus NMI. You need to take
this into account in the return value of the function.

This version fixes initial patch which was missing changes to
perf_event_intel_ds.c.

Signed-off-by: Stephane Eranian <eranian@google.com>
Acked-by: Don Zickus <dzickus@redhat.com>
Cc: peterz@infradead.org
Cc: paulus@samba.org
Cc: davem@davemloft.net
Cc: fweisbec@gmail.com
Cc: perfmon2-devel@lists.sf.net
Cc: eranian@gmail.com
Cc: robert.richter@amd.com
LKML-Reference: <4c8a1686.aae9d80a.5aa4.5e35@mx.google.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoLinux 2.6.36-rc4
Linus Torvalds [Sun, 12 Sep 2010 23:07:37 +0000 (16:07 -0700)]
Linux 2.6.36-rc4

14 years agodocbook: skip files with no docs since they generate scary warnings
Randy Dunlap [Sat, 11 Sep 2010 22:55:26 +0000 (15:55 -0700)]
docbook: skip files with no docs since they generate scary warnings

Fix docbook templates that reference files that do not contain the
expected kernel-doc notation.

Fixes these warnings:

  Warning(arch/x86/include/asm/unaligned.h): no structured comments found
  Warning(lib/vsprintf.c): no structured comments found

These cause errors in the generated html output, like below, so drop
these lines.

  Name
  arch/x86/include/asm/unaligned.h - Document generation inconsistency
  Oops
  Warning
  The template for this document tried to insert the structured comment from the file arch/x86/include/asm/unaligned.h at this point, but none was found. This dummy section is inserted to allow generation to continue.

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agodocbook: warn on unused doc entries
Johannes Berg [Sat, 11 Sep 2010 22:55:22 +0000 (15:55 -0700)]
docbook: warn on unused doc entries

When you don't use !E or !I but only !F, then it's very easy to miss
including some functions, structs etc.  in documentation.  To help
finding which ones were missed, allow printing out the unused ones as
warnings.

For example, using this on mac80211 yields a lot of warnings like this:

  Warning: didn't use docs for DOC: mac80211 workqueue
  Warning: didn't use docs for ieee80211_max_queues
  Warning: didn't use docs for ieee80211_bss_change
  Warning: didn't use docs for ieee80211_bss_conf

when generating the documentation for it.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agokernel-doc: ignore case when stripping attributes
Johannes Berg [Sat, 11 Sep 2010 22:55:12 +0000 (15:55 -0700)]
kernel-doc: ignore case when stripping attributes

There are valid attributes that could have upper case letters, but we
still want to remove, like for example
__attribute__((aligned(NETDEV_ALIGN)))
as encountered in the wireless code.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoMerge branch 'pm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/suspe...
Linus Torvalds [Sat, 11 Sep 2010 22:50:53 +0000 (15:50 -0700)]
Merge branch 'pm-fixes' of git://git./linux/kernel/git/rafael/suspend-2.6

* 'pm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/suspend-2.6:
  PM / Hibernate: Avoid hitting OOM during preallocation of memory
  PM QoS: Correct pr_debug() misuse and improve parameter checks
  PM: Prevent waiting forever on asynchronous resume after failing suspend

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6
Linus Torvalds [Sat, 11 Sep 2010 19:17:02 +0000 (12:17 -0700)]
Merge git://git./linux/kernel/git/jejb/scsi-rc-fixes-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6:
  [SCSI] fix use-after-free in scsi_init_io()
  [SCSI] sd: fix medium-removal bug
  [SCSI] qla2xxx: Update version number to 8.03.04-k0.
  [SCSI] qla2xxx: Check for empty slot in request queue before posting Command type 6 request.
  [SCSI] qla2xxx: Cover UNDERRUN case where SCSI status is set.
  [SCSI] qla2xxx: Correctly set fw hung and complete only waiting mbx.
  [SCSI] qla2xxx: Reset seconds_since_last_heartbeat correctly.
  [SCSI] qla2xxx: make rport deletions explicit during vport removal
  [SCSI] qla2xxx: Fix vport delete issues
  [SCSI] sd, sym53c8xx: Remove warnings after vsprintf %pV introducation.
  [SCSI] Fix warning: zero-length gnu_printf format string
  [SCSI] hpsa: disable doorbell reset on reset_devices
  [SCSI] be2iscsi: Fix for Login failure
  [SCSI] fix bio.bi_rw handling

14 years agoPM / Hibernate: Avoid hitting OOM during preallocation of memory
Rafael J. Wysocki [Sat, 11 Sep 2010 18:58:27 +0000 (20:58 +0200)]
PM / Hibernate: Avoid hitting OOM during preallocation of memory

There is a problem in hibernate_preallocate_memory() that it calls
preallocate_image_memory() with an argument that may be greater than
the total number of available non-highmem memory pages.  If that's
the case, the OOM condition is guaranteed to trigger, which in turn
can cause significant slowdown to occur during hibernation.

To avoid that, make preallocate_image_memory() adjust its argument
before calling preallocate_image_pages(), so that the total number of
saveable non-highem pages left is not less than the minimum size of
a hibernation image.  Change hibernate_preallocate_memory() to try to
allocate from highmem if the number of pages allocated by
preallocate_image_memory() is too low.

Modify free_unnecessary_pages() to take all possible memory
allocation patterns into account.

Reported-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Tested-by: M. Vefa Bicakci <bicave@superonline.com>
14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Sat, 11 Sep 2010 15:06:38 +0000 (08:06 -0700)]
Merge git://git./linux/kernel/git/davem/net-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (28 commits)
  ipheth: remove incorrect devtype to WWAN
  MAINTAINERS: Add CAIF
  sctp: fix test for end of loop
  KS8851: Correct RX packet allocation
  udp: add rehash on connect()
  net: blackhole route should always be recalculated
  ipv4: Suppress lockdep-RCU false positive in FIB trie (3)
  niu: Fix kernel buffer overflow for ETHTOOL_GRXCLSRLALL
  ipvs: fix active FTP
  gro: Re-fix different skb headrooms
  via-velocity: Turn scatter-gather support back off.
  ipv4: Fix reverse path filtering with multipath routing.
  UNIX: Do not loop forever at unix_autobind().
  PATCH: b44 Handle RX FIFO overflow better (simplified)
  irda: off by one
  3c59x: Fix deadlock in vortex_error()
  netfilter: discard overlapping IPv6 fragment
  ipv6: discard overlapping fragment
  net: fix tx queue selection for bridged devices implementing select_queue
  bonding: Fix jiffies overflow problems (again)
  ...

Fix up trivial conflicts due to the same cgroup API thinko fix going
through both Andrew and the networking tree.  However, there were small
differences between the two, with Andrew's version generally being the
nicer one, and the one I merged first. So pick that one.

Conflicts in: include/linux/cgroup.h and kernel/cgroup.c

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6
Linus Torvalds [Sat, 11 Sep 2010 15:01:09 +0000 (08:01 -0700)]
Merge git://git./linux/kernel/git/davem/sparc-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6:
  sparc: Kill all BKL usage.

14 years agoMerge branch 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 11 Sep 2010 14:59:49 +0000 (07:59 -0700)]
Merge branch 'sched-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  x86, tsc: Fix a preemption leak in restore_sched_clock_state()
  sched: Move sched_avg_update() to update_cpu_load()

14 years agox86, tsc: Fix a preemption leak in restore_sched_clock_state()
Peter Zijlstra [Fri, 10 Sep 2010 20:32:53 +0000 (22:32 +0200)]
x86, tsc: Fix a preemption leak in restore_sched_clock_state()

Doh, a real life genuine preemption leak..

This caused a suspend failure.

Reported-bisected-and-tested-by-the-invaluable: Jeff Chua <jeff.chua.linux@gmail.com>
Acked-by: Suresh Siddha <suresh.b.siddha@intel.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Rafael J. Wysocki <rjw@sisk.pl>
Cc: Nico Schottelius <nico-linux-20100709@schottelius.org>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Florian Pritz <flo@xssn.at>
Cc: Suresh Siddha <suresh.b.siddha@intel.com>
Cc: Len Brown <lenb@kernel.org>
Cc: <stable@kernel.org> # Greg, please apply after: cd7240c ("x86, tsc, sched: Recompute cyc2ns_offset's during resume from")
sleep states
LKML-Reference: <1284150773.402.122.camel@laptop>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoMerge branch 'drm-intel-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/ickle...
Linus Torvalds [Sat, 11 Sep 2010 01:19:43 +0000 (18:19 -0700)]
Merge branch 'drm-intel-fixes' of git://git./linux/kernel/git/ickle/drm-intel

* 'drm-intel-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/ickle/drm-intel:
  drm/i915: don't enable self-refresh on Ironlake
  drm/i915: Double check that the wait_request is not pending before warning
  Revert "drm/i915: Warn if we run out of FIFO space for a mode"
  Revert "drm/i915: Allow LVDS on pipe A on gen4+"
  Revert "drm/i915: Enable RC6 on Ironlake."

14 years agoMerge branch 'for-linus' of git://oss.sgi.com/xfs/xfs
Linus Torvalds [Sat, 11 Sep 2010 01:19:26 +0000 (18:19 -0700)]
Merge branch 'for-linus' of git://oss.sgi.com/xfs/xfs

* 'for-linus' of git://oss.sgi.com/xfs/xfs:
  xfs: log IO completion workqueue is a high priority queue
  xfs: prevent reading uninitialized stack memory

14 years agox86, tsc: Fix a preemption leak in restore_sched_clock_state()
Peter Zijlstra [Fri, 10 Sep 2010 20:32:53 +0000 (22:32 +0200)]
x86, tsc: Fix a preemption leak in restore_sched_clock_state()

A real life genuine preemption leak..

Reported-and-tested-by: Jeff Chua <jeff.chua.linux@gmail.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Suresh Siddha <suresh.b.siddha@intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoPM QoS: Correct pr_debug() misuse and improve parameter checks
mark gross [Thu, 9 Sep 2010 21:20:09 +0000 (23:20 +0200)]
PM QoS: Correct pr_debug() misuse and improve parameter checks

Correct some pr_debug() misuse and add a stronger parameter check to
pm_qos_write() for the ASCII hex value case.  Thanks to Dan Carpenter
for pointing out the problem!

Signed-off-by: mark gross <markgross@thegnar.org>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
14 years agoperf: Fix perf_init_event()
Peter Zijlstra [Fri, 10 Sep 2010 15:38:06 +0000 (17:38 +0200)]
perf: Fix perf_init_event()

We ought to return -ENOENT when non of the registered PMUs
recognise the requested event.

This fixes a boot crash that occurs if no PMU is available
but the NMI watchdog tries to register an event.

Reported-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoxfs: log IO completion workqueue is a high priority queue
Dave Chinner [Wed, 8 Sep 2010 09:00:22 +0000 (09:00 +0000)]
xfs: log IO completion workqueue is a high priority queue

The workqueue implementation in 2.6.36-rcX has changed, resulting
in the workqueues no longer having dedicated threads for work
processing. This has caused severe livelocks under heavy parallel
create workloads because the log IO completions have been getting
held up behind metadata IO completions.  Hence log commits would
stall, memory allocation would stall because pages could not be
cleaned, and lock contention on the AIL during inode IO completion
processing was being seen to slow everything down even further.

By making the log Io completion workqueue a high priority workqueue,
they are queued ahead of all data/metadata IO completions and
processed before the data/metadata completions. Hence the log never
gets stalled, and operations needed to clean memory can continue as
quickly as possible. This avoids the livelock conditions and allos
the system to keep running under heavy load as per normal.

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Alex Elder <aelder@sgi.com>
14 years agoexecve: make responsive to SIGKILL with large arguments
Roland McGrath [Wed, 8 Sep 2010 02:37:06 +0000 (19:37 -0700)]
execve: make responsive to SIGKILL with large arguments

An execve with a very large total of argument/environment strings
can take a really long time in the execve system call.  It runs
uninterruptibly to count and copy all the strings.  This change
makes it abort the exec quickly if sent a SIGKILL.

Note that this is the conservative change, to interrupt only for
SIGKILL, by using fatal_signal_pending().  It would be perfectly
correct semantics to let any signal interrupt the string-copying in
execve, i.e. use signal_pending() instead of fatal_signal_pending().
We'll save that change for later, since it could have user-visible
consequences, such as having a timer set too quickly make it so that
an execve can never complete, though it always happened to work before.

Signed-off-by: Roland McGrath <roland@redhat.com>
Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoexecve: improve interactivity with large arguments
Roland McGrath [Wed, 8 Sep 2010 02:36:28 +0000 (19:36 -0700)]
execve: improve interactivity with large arguments

This adds a preemption point during the copying of the argument and
environment strings for execve, in copy_strings().  There is already
a preemption point in the count() loop, so this doesn't add any new
points in the abstract sense.

When the total argument+environment strings are very large, the time
spent copying them can be much more than a normal user time slice.
So this change improves the interactivity of the rest of the system
when one process is doing an execve with very large arguments.

Signed-off-by: Roland McGrath <roland@redhat.com>
Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agosetup_arg_pages: diagnose excessive argument size
Roland McGrath [Wed, 8 Sep 2010 02:35:49 +0000 (19:35 -0700)]
setup_arg_pages: diagnose excessive argument size

The CONFIG_STACK_GROWSDOWN variant of setup_arg_pages() does not
check the size of the argument/environment area on the stack.
When it is unworkably large, shift_arg_pages() hits its BUG_ON.
This is exploitable with a very large RLIMIT_STACK limit, to
create a crash pretty easily.

Check that the initial stack is not too large to make it possible
to map in any executable.  We're not checking that the actual
executable (or intepreter, for binfmt_elf) will fit.  So those
mappings might clobber part of the initial stack mapping.  But
that is just userland lossage that userland made happen, not a
kernel problem.

Signed-off-by: Roland McGrath <roland@redhat.com>
Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoMerge branch 'kvm-updates/2.6.36' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Linus Torvalds [Fri, 10 Sep 2010 15:02:45 +0000 (08:02 -0700)]
Merge branch 'kvm-updates/2.6.36' of git://git./virt/kvm/kvm

* 'kvm-updates/2.6.36' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
  KVM: x86: Perform hardware_enable in CPU_STARTING callback
  KVM: i8259: fix migration
  KVM: fix i8259 oops when no vcpus are online
  KVM: x86 emulator: fix regression with cmpxchg8b on i386 hosts

14 years agoMerge branch 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 10 Sep 2010 14:31:24 +0000 (07:31 -0700)]
Merge branch 'perf-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  tracing: t_start: reset FTRACE_ITER_HASH in case of seek/pread
  perf symbols: Fix multiple initialization of symbol system
  perf: Fix CPU hotplug
  perf, trace: Fix module leak
  tracing/kprobe: Fix handling of C-unlike argument names
  tracing/kprobes: Fix handling of argument names
  perf probe: Fix handling of arguments names
  perf probe: Fix return probe support
  tracing/kprobe: Fix a memory leak in error case
  tracing: Do not allow llseek to set_ftrace_filter

14 years agoKEYS: Fix bug in keyctl_session_to_parent() if parent has no session keyring
David Howells [Fri, 10 Sep 2010 08:59:51 +0000 (09:59 +0100)]
KEYS: Fix bug in keyctl_session_to_parent() if parent has no session keyring

Fix a bug in keyctl_session_to_parent() whereby it tries to check the ownership
of the parent process's session keyring whether or not the parent has a session
keyring [CVE-2010-2960].

This results in the following oops:

  BUG: unable to handle kernel NULL pointer dereference at 00000000000000a0
  IP: [<ffffffff811ae4dd>] keyctl_session_to_parent+0x251/0x443
  ...
  Call Trace:
   [<ffffffff811ae2f3>] ? keyctl_session_to_parent+0x67/0x443
   [<ffffffff8109d286>] ? __do_fault+0x24b/0x3d0
   [<ffffffff811af98c>] sys_keyctl+0xb4/0xb8
   [<ffffffff81001eab>] system_call_fastpath+0x16/0x1b

if the parent process has no session keyring.

If the system is using pam_keyinit then it mostly protected against this as all
processes derived from a login will have inherited the session keyring created
by pam_keyinit during the log in procedure.

To test this, pam_keyinit calls need to be commented out in /etc/pam.d/.

Reported-by: Tavis Ormandy <taviso@cmpxchg8b.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: Tavis Ormandy <taviso@cmpxchg8b.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoKEYS: Fix RCU no-lock warning in keyctl_session_to_parent()
David Howells [Fri, 10 Sep 2010 08:59:46 +0000 (09:59 +0100)]
KEYS: Fix RCU no-lock warning in keyctl_session_to_parent()

There's an protected access to the parent process's credentials in the middle
of keyctl_session_to_parent().  This results in the following RCU warning:

  ===================================================
  [ INFO: suspicious rcu_dereference_check() usage. ]
  ---------------------------------------------------
  security/keys/keyctl.c:1291 invoked rcu_dereference_check() without protection!

  other info that might help us debug this:

  rcu_scheduler_active = 1, debug_locks = 0
  1 lock held by keyctl-session-/2137:
   #0:  (tasklist_lock){.+.+..}, at: [<ffffffff811ae2ec>] keyctl_session_to_parent+0x60/0x236

  stack backtrace:
  Pid: 2137, comm: keyctl-session- Not tainted 2.6.36-rc2-cachefs+ #1
  Call Trace:
   [<ffffffff8105606a>] lockdep_rcu_dereference+0xaa/0xb3
   [<ffffffff811ae379>] keyctl_session_to_parent+0xed/0x236
   [<ffffffff811af77e>] sys_keyctl+0xb4/0xb6
   [<ffffffff81001eab>] system_call_fastpath+0x16/0x1b

The code should take the RCU read lock to make sure the parents credentials
don't go away, even though it's holding a spinlock and has IRQ disabled.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoMerge branch 'for-linus' of git://git.kernel.dk/linux-2.6-block
Linus Torvalds [Fri, 10 Sep 2010 14:26:27 +0000 (07:26 -0700)]
Merge branch 'for-linus' of git://git.kernel.dk/linux-2.6-block

* 'for-linus' of git://git.kernel.dk/linux-2.6-block:
  block: Range check cpu in blk_cpu_to_group
  scatterlist: prevent invalid free when alloc fails
  writeback: Fix lost wake-up shutting down writeback thread
  writeback: do not lose wakeup events when forking bdi threads
  cciss: fix reporting of max queue depth since init
  block: switch s390 tape_block and mg_disk to elevator_change()
  block: add function call to switch the IO scheduler from a driver
  fs/bio-integrity.c: return -ENOMEM on kmalloc failure
  bio-integrity.c: remove dependency on __GFP_NOFAIL
  BLOCK: fix bio.bi_rw handling
  block: put dev->kobj in blk_register_queue fail path
  cciss: handle allocation failure
  cfq-iosched: Documentation help for new tunables
  cfq-iosched: blktrace print per slice sector stats
  cfq-iosched: Implement tunable group_idle
  cfq-iosched: Do group share accounting in IOPS when slice_idle=0
  cfq-iosched: Do not idle if slice_idle=0
  cciss: disable doorbell reset on reset_devices
  blkio: Fix return code for mkdir calls

14 years agoMerge branch 'at91-fixes-for-linus' of git://github.com/at91linux/linux-2.6-at91
Linus Torvalds [Fri, 10 Sep 2010 14:24:51 +0000 (07:24 -0700)]
Merge branch 'at91-fixes-for-linus' of git://github.com/at91linux/linux-2.6-at91

* 'at91-fixes-for-linus' of git://github.com/at91linux/linux-2.6-at91:
  AT91: at91sam9261ek: remove C99 comments but keep information
  AT91: at91sam9261ek board: remove warnings related to use of SPI or SD/MMC
  AT91: dm9000 initialization update
  AT91: SAM9G45 - add a separate clock entry for every single TC block
  AT91: clock: peripheral clocks can have other parent than mck
  AT91: change dma resource index

14 years agoperf: Ensure we call add_event_to_ctx() with the right locks held
Peter Zijlstra [Fri, 10 Sep 2010 10:51:54 +0000 (12:51 +0200)]
perf: Ensure we call add_event_to_ctx() with the right locks held

Even though we call it from the inherit path, where the child is
not yet accessible, we need to hold ctx->lock, add_event_to_ctx()
assumes IRQs are disabled.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6
Linus Torvalds [Fri, 10 Sep 2010 14:23:45 +0000 (07:23 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/tiwai/sound-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6:
  ALSA: rawmidi: fix the get next midi device ioctl
  ALSA: hda - Fix wrong HP pin detection in snd_hda_parse_pin_def_config()
  ALSA: seq/oss - Fix double-free at error path of snd_seq_oss_open()
  ALSA: msnd-classic: Fix invalid cfg parameter
  ALSA: hda - Enable PC-beep for EeePC with ALC269 codec
  ALSA: hda - Add errata initverb sequence for CS42xx codecs
  ALSA: usb - Release capture substream URBs properly
  ALSA: virtuoso: fix setting of Xonar DS line-in/mic-in controls
  ALSA: virtuoso: work around missing reset in the Xonar DS Windows driver
  ALSA: hda - Add quirk for Lenovo T400s
  ALSA: usb-audio: fix detection of vendor-specific device protocol settings
  ALSA: usb-audio: Assume first control interface is for audio
  ALSA: hda - Add a new hp-laptop model for Conexant 5066, tested on HP G60

14 years agodrm/i915: don't enable self-refresh on Ironlake
Jesse Barnes [Thu, 9 Sep 2010 18:58:02 +0000 (11:58 -0700)]
drm/i915: don't enable self-refresh on Ironlake

We don't know how to enable it safely, especially as outputs turn on and
off.  When disabling LP1 we also need to make sure LP2 and 3 are already
disabled.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=29173
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=29082
Reported-by: Chris Lord <chris@linux.intel.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Tested-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: stable@kernel.org
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agoxfs: prevent reading uninitialized stack memory
Dan Rosenberg [Mon, 6 Sep 2010 22:24:57 +0000 (18:24 -0400)]
xfs: prevent reading uninitialized stack memory

The XFS_IOC_FSGETXATTR ioctl allows unprivileged users to read 12
bytes of uninitialized stack memory, because the fsxattr struct
declared on the stack in xfs_ioc_fsgetxattr() does not alter (or zero)
the 12-byte fsx_pad member before copying it back to the user.  This
patch takes care of it.

Signed-off-by: Dan Rosenberg <dan.j.rosenberg@gmail.com>
Reviewed-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
14 years agoAT91: at91sam9261ek: remove C99 comments but keep information
Nicolas Ferre [Fri, 10 Sep 2010 12:36:06 +0000 (14:36 +0200)]
AT91: at91sam9261ek: remove C99 comments but keep information

Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
14 years agoAT91: at91sam9261ek board: remove warnings related to use of SPI or SD/MMC
Nicolas Ferre [Fri, 10 Sep 2010 09:26:42 +0000 (11:26 +0200)]
AT91: at91sam9261ek board: remove warnings related to use of SPI or SD/MMC

The sd/mmc data structure is not used if SPI is selected. The configuration
of PIO on the board prevent from using both interfaces at the same time
(board dependent).
Remove the warnings at compilation time adding a preprocessor condition.

Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
14 years agoAT91: dm9000 initialization update
Nicolas Ferre [Fri, 10 Sep 2010 09:38:43 +0000 (11:38 +0200)]
AT91: dm9000 initialization update

Add information in dm9000 mac/phy chip initialization:
- irq resource details
- platform data details

Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
14 years agoblock: Range check cpu in blk_cpu_to_group
Brian King [Fri, 10 Sep 2010 07:03:21 +0000 (09:03 +0200)]
block: Range check cpu in blk_cpu_to_group

While testing CPU DLPAR, the following problem was discovered.
We were DLPAR removing the first CPU, which in this case was
logical CPUs 0-3. CPUs 0-2 were already marked offline and
we were in the process of offlining CPU 3. After marking
the CPU inactive and offline in cpu_disable, but before the
cpu was completely idle (cpu_die), we ended up in __make_request
on CPU 3. There we looked at the topology map to see which CPU
to complete the I/O on and found no CPUs in the cpu_sibling_map.
This resulted in the block layer setting the completion cpu
to be NR_CPUS, which then caused an oops when we tried to
complete the I/O.

Fix this by sanity checking the value we return from blk_cpu_to_group
to be a valid cpu value.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
14 years agoMerge branch 'fix/hda' into for-linus
Takashi Iwai [Fri, 10 Sep 2010 06:27:00 +0000 (08:27 +0200)]
Merge branch 'fix/hda' into for-linus

14 years agoMerge branch 'tip/perf/urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/roste...
Ingo Molnar [Fri, 10 Sep 2010 06:05:34 +0000 (08:05 +0200)]
Merge branch 'tip/perf/urgent' of git://git./linux/kernel/git/rostedt/linux-2.6-trace into perf/urgent

14 years agoMerge branch 'perf/urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/acme...
Ingo Molnar [Fri, 10 Sep 2010 05:34:14 +0000 (07:34 +0200)]
Merge branch 'perf/urgent' of git://git./linux/kernel/git/acme/linux-2.6 into perf/urgent

14 years agoMerge branch 'vhost-net' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost
David S. Miller [Fri, 10 Sep 2010 04:59:51 +0000 (21:59 -0700)]
Merge branch 'vhost-net' of git://git./linux/kernel/git/mst/vhost

14 years agoipheth: remove incorrect devtype to WWAN
Dan Williams [Wed, 8 Sep 2010 07:50:47 +0000 (07:50 +0000)]
ipheth: remove incorrect devtype to WWAN

The 'wwan' devtype is meant for devices that require preconfiguration
and *every* time setup before the ethernet interface can be used, like
cellular modems which require a series of setup commands on serial ports
or other mechanisms before the ethernet interface will handle packets.

As ipheth only requires one-per-hotplug pairing setup with no
preconfiguration (like APN, phone #, etc) and the network interface is
usable at any time after that initial setup, remove the incorrect
devtype wwan.

Signed-off-by: Dan Williams <dcbw@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agoMAINTAINERS: Add CAIF
Joe Perches [Tue, 7 Sep 2010 20:33:24 +0000 (20:33 +0000)]
MAINTAINERS: Add CAIF

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agoMerge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzi...
Linus Torvalds [Fri, 10 Sep 2010 03:28:19 +0000 (20:28 -0700)]
Merge branch 'upstream-linus' of git://git./linux/kernel/git/jgarzik/libata-dev

* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev:
  libata-sff: Reenable Port Multiplier after libata-sff remodeling.
  libata: skip EH autopsy and recovery during suspend
  ahci: AHCI and RAID mode SATA patch for Intel Patsburg DeviceIDs
  ata_piix: IDE Mode SATA patch for Intel Patsburg DeviceIDs
  libata,pata_via: revert ata_wait_idle() removal from ata_sff/via_tf_load()
  ahci: fix hang on failed softreset
  pata_artop: Fix device ID parity check

14 years agotracing: t_start: reset FTRACE_ITER_HASH in case of seek/pread
Chris Wright [Thu, 9 Sep 2010 23:34:59 +0000 (16:34 -0700)]
tracing: t_start: reset FTRACE_ITER_HASH in case of seek/pread

Be sure to avoid entering t_show() with FTRACE_ITER_HASH set without
having properly started the iterator to iterate the hash.  This case is
degenerate and, as discovered by Robert Swiecki, can cause t_hash_show()
to misuse a pointer.  This causes a NULL ptr deref with possible security
implications.  Tracked as CVE-2010-3079.

Cc: Robert Swiecki <swiecki@google.com>
Cc: Eugene Teo <eugene@redhat.com>
Cc: <stable@kernel.org>
Signed-off-by: Chris Wright <chrisw@sous-sol.org>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
14 years agolibata-sff: Reenable Port Multiplier after libata-sff remodeling.
Gwendal Grignou [Tue, 31 Aug 2010 23:20:36 +0000 (16:20 -0700)]
libata-sff: Reenable Port Multiplier after libata-sff remodeling.

Keep track of the link on the which the current request is in progress.
It allows support of links behind port multiplier.

Not all libata-sff is PMP compliant. Code for native BMDMA controller
does not take in accound PMP.

Tested on Marvell 7042 and Sil7526.

Signed-off-by: Gwendal Grignou <gwendal@google.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
14 years agolibata: skip EH autopsy and recovery during suspend
Tejun Heo [Tue, 7 Sep 2010 12:05:31 +0000 (14:05 +0200)]
libata: skip EH autopsy and recovery during suspend

For some mysterious reason, certain hardware reacts badly to usual EH
actions while the system is going for suspend.  As the devices won't
be needed until the system is resumed, ask EH to skip usual autopsy
and recovery and proceed directly to suspend.

Signed-off-by: Tejun Heo <tj@kernel.org>
Tested-by: Stephan Diestelhorst <stephan.diestelhorst@amd.com>
Cc: stable@kernel.org
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
14 years agoahci: AHCI and RAID mode SATA patch for Intel Patsburg DeviceIDs
Seth Heasley [Thu, 9 Sep 2010 16:44:56 +0000 (09:44 -0700)]
ahci: AHCI and RAID mode SATA patch for Intel Patsburg DeviceIDs

This patch adds the Intel Patsburg (PCH) SATA AHCI and RAID Controller
DeviceIDs.

Signed-off-by: Seth Heasley <seth.heasley@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
14 years agoata_piix: IDE Mode SATA patch for Intel Patsburg DeviceIDs
Seth Heasley [Thu, 9 Sep 2010 16:42:40 +0000 (09:42 -0700)]
ata_piix: IDE Mode SATA patch for Intel Patsburg DeviceIDs

This patch adds the Intel Patsburg (PCH) IDE mode SATA Controller DeviceIDs.

Signed-off-by: Seth Heasley <seth.heasley@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
14 years agolibata,pata_via: revert ata_wait_idle() removal from ata_sff/via_tf_load()
Tejun Heo [Thu, 9 Sep 2010 15:13:31 +0000 (17:13 +0200)]
libata,pata_via: revert ata_wait_idle() removal from ata_sff/via_tf_load()

Commit 978c0666 (libata: Remove excess delay in the tf_load path)
removed ata_wait_idle() from ata_sff_tf_load() and via_tf_load().
This caused obscure detection problems in sata_sil.

  https://bugzilla.kernel.org/show_bug.cgi?id=16606

The commit was pure performance optimization.  Revert it for now.

Reported-by: Dieter Plaetinck <dieter@plaetinck.be>
Reported-by: Jan Beulich <JBeulich@novell.com>
Bisected-by: gianluca <gianluca@sottospazio.it>
Cc: stable@kernel.org
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
14 years agominix: fix regression in minix_mkdir()
Jorge Boncompte [DTI2] [Thu, 9 Sep 2010 23:38:19 +0000 (16:38 -0700)]
minix: fix regression in minix_mkdir()

Commit 9eed1fb721c ("minix: replace inode uid,gid,mode init with helper")
broke directory creation on minix filesystems.

Fix it by passing the needed mode flag to inode init helper.

Signed-off-by: Jorge Boncompte [DTI2] <jorge@dti2.net>
Cc: Dmitry Monakhov <dmonakhov@openvz.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: <stable@kernel.org> [2.6.35.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agomm: page allocator: drain per-cpu lists after direct reclaim allocation fails
Mel Gorman [Thu, 9 Sep 2010 23:38:18 +0000 (16:38 -0700)]
mm: page allocator: drain per-cpu lists after direct reclaim allocation fails

When under significant memory pressure, a process enters direct reclaim
and immediately afterwards tries to allocate a page.  If it fails and no
further progress is made, it's possible the system will go OOM.  However,
on systems with large amounts of memory, it's possible that a significant
number of pages are on per-cpu lists and inaccessible to the calling
process.  This leads to a process entering direct reclaim more often than
it should increasing the pressure on the system and compounding the
problem.

This patch notes that if direct reclaim is making progress but allocations
are still failing that the system is already under heavy pressure.  In
this case, it drains the per-cpu lists and tries the allocation a second
time before continuing.

Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Reviewed-by: Minchan Kim <minchan.kim@gmail.com>
Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Reviewed-by: Christoph Lameter <cl@linux.com>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agomm: page allocator: calculate a better estimate of NR_FREE_PAGES when memory is low...
Christoph Lameter [Thu, 9 Sep 2010 23:38:17 +0000 (16:38 -0700)]
mm: page allocator: calculate a better estimate of NR_FREE_PAGES when memory is low and kswapd is awake

Ordinarily watermark checks are based on the vmstat NR_FREE_PAGES as it is
cheaper than scanning a number of lists.  To avoid synchronization
overhead, counter deltas are maintained on a per-cpu basis and drained
both periodically and when the delta is above a threshold.  On large CPU
systems, the difference between the estimated and real value of
NR_FREE_PAGES can be very high.  If NR_FREE_PAGES is much higher than
number of real free page in buddy, the VM can allocate pages below min
watermark, at worst reducing the real number of pages to zero.  Even if
the OOM killer kills some victim for freeing memory, it may not free
memory if the exit path requires a new page resulting in livelock.

This patch introduces a zone_page_state_snapshot() function (courtesy of
Christoph) that takes a slightly more accurate view of an arbitrary vmstat
counter.  It is used to read NR_FREE_PAGES while kswapd is awake to avoid
the watermark being accidentally broken.  The estimate is not perfect and
may result in cache line bounces but is expected to be lighter than the
IPI calls necessary to continually drain the per-cpu counters while kswapd
is awake.

Signed-off-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agomm: page allocator: update free page counters after pages are placed on the free...
Mel Gorman [Thu, 9 Sep 2010 23:38:16 +0000 (16:38 -0700)]
mm: page allocator: update free page counters after pages are placed on the free list

When allocating a page, the system uses NR_FREE_PAGES counters to
determine if watermarks would remain intact after the allocation was made.
This check is made without interrupts disabled or the zone lock held and
so is race-prone by nature.  Unfortunately, when pages are being freed in
batch, the counters are updated before the pages are added on the list.
During this window, the counters are misleading as the pages do not exist
yet.  When under significant pressure on systems with large numbers of
CPUs, it's possible for processes to make progress even though they should
have been stalled.  This is particularly problematic if a number of the
processes are using GFP_ATOMIC as the min watermark can be accidentally
breached and in extreme cases, the system can livelock.

This patch updates the counters after the pages have been added to the
list.  This makes the allocator more cautious with respect to preserving
the watermarks and mitigates livelock possibilities.

[akpm@linux-foundation.org: avoid modifying incoming args]
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Reviewed-by: Rik van Riel <riel@redhat.com>
Reviewed-by: Minchan Kim <minchan.kim@gmail.com>
Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Reviewed-by: Christoph Lameter <cl@linux.com>
Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agovmstat: update zone stat threshold when onlining a cpu
KAMEZAWA Hiroyuki [Thu, 9 Sep 2010 23:38:14 +0000 (16:38 -0700)]
vmstat: update zone stat threshold when onlining a cpu

refresh_zone_stat_thresholds() calculates parameter based on the number of
online cpus.  It's called at cpu offlining but needs to be called at
onlining, too.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Christoph Lameter <cl@linux-foundation.org>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agovfs: take O_NONBLOCK out of the O_* uniqueness test
James Bottomley [Thu, 9 Sep 2010 23:38:12 +0000 (16:38 -0700)]
vfs: take O_NONBLOCK out of the O_* uniqueness test

O_NONBLOCK on parisc has a dual value:

#define O_NONBLOCK 000200004 /* HPUX has separate NDELAY & NONBLOCK */

It is caught by the O_* bits uniqueness check and leads to a parisc
compile error.  The fix would be to take O_NONBLOCK out.

Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Cc: Jamie Lokier <jamie@shareable.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoswap: discard while swapping only if SWAP_FLAG_DISCARD
Hugh Dickins [Thu, 9 Sep 2010 23:38:11 +0000 (16:38 -0700)]
swap: discard while swapping only if SWAP_FLAG_DISCARD

Tests with recent firmware on Intel X25-M 80GB and OCZ Vertex 60GB SSDs
show a shift since I last tested in December: in part because of firmware
updates, in part because of the necessary move from barriers to awaiting
completion at the block layer.  While discard at swapon still shows as
slightly beneficial on both, discarding 1MB swap cluster when allocating
is now disadvanteous: adds 25% overhead on Intel, adds 230% on OCZ (YMMV).

Surrender: discard as presently implemented is more hindrance than help
for swap; but might prove useful on other devices, or with improvements.
So continue to do the discard at swapon, but make discard while swapping
conditional on a SWAP_FLAG_DISCARD to sys_swapon() (which has been using
only the lower 16 bits of int flags).

We can add a --discard or -d to swapon(8), and a "discard" to swap in
/etc/fstab: matching the mount option for btrfs, ext4, fat, gfs2, nilfs2.

Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Nigel Cunningham <nigel@tuxonice.net>
Cc: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <jaxboe@fusionio.com>
Cc: James Bottomley <James.Bottomley@hansenpartnership.com>
Cc: "Martin K. Petersen" <martin.petersen@oracle.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoswap: do not send discards as barriers
Christoph Hellwig [Thu, 9 Sep 2010 23:38:10 +0000 (16:38 -0700)]
swap: do not send discards as barriers

The swap code already uses synchronous discards, no need to add I/O
barriers.

This fixes the worst of the terrible slowdown in swap allocation for
hibernation, reported on 2.6.35 by Nigel Cunningham; but does not entirely
eliminate that regression.

[tj@kernel.org: superflous newlines removed]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Tested-by: Nigel Cunningham <nigel@tuxonice.net>
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: Jens Axboe <jaxboe@fusionio.com>
Cc: James Bottomley <James.Bottomley@hansenpartnership.com>
Cc: "Martin K. Petersen" <martin.petersen@oracle.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoswap: prevent reuse during hibernation
Hugh Dickins [Thu, 9 Sep 2010 23:38:09 +0000 (16:38 -0700)]
swap: prevent reuse during hibernation

Move the hibernation check from scan_swap_map() into try_to_free_swap():
to catch not only the common case when hibernation's allocation itself
triggers swap reuse, but also the less likely case when concurrent page
reclaim (shrink_page_list) might happen to try_to_free_swap from a page.

Hibernation already clears __GFP_IO from the gfp_allowed_mask, to stop
reclaim from going to swap: check that to prevent swap reuse too.

Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: Ondrej Zary <linux@rainbow-software.org>
Cc: Andrea Gelmini <andrea.gelmini@gmail.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Nigel Cunningham <nigel@tuxonice.net>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>